Transcriptional gene silencing by hypermethylation of CpG islands spanning the promoter regions of genes is a common and important mechanism in carcinogenesis. Hypermethylation leads to inactivation of gene expression, and this epigenetic alteration is considered a key mechanism for long-term gene silencing. Despite of thousands of reports in the literature describing hypermethylation of specific genes in almost every type of human cancer, the mechanisms of CpG island hypermethylation have remained obscure. This application will focus on mechanistic studies that will investigate the molecular pathways leading to DNA methylation changes in tumors with emphasis on DNA hypermethylation. As a key technology to accomplish these goals, we have recently developed a method that enables the precise genome-wide analysis of DNA cytosine methylation patterns in mammalian DNA. Using genome-wide methylation profiling, we will analyze the extent by which chemical carcinogens can induce DNA methylation changes. Using cell culture models, we will expose non-transformed human cells (fibroblasts, bronchial epithelial cells and mammary epithelial cells) to several chemical carcinogens. We will analyze DNA methylation changes within 27,800 CpG islands and along the entire short arms of chromosomes 3, 7, and 9. DNA methylation changes will be analyzed in a mouse skin carcinogenesis model in order to determine if epigenetic changes arise during the initiation or promotion of stages of carcinogenesis. We will also analyze if the same carcinogens can induce changes in histone modification patterns that may subsequently predispose the CpG islands to de novo DNA methylation. In another Specific Aim, we will analyze the hypothesis that DNA methylation changes can be induced by endogenous mechanisms including inflammation and oxidative DNA damage. We will investigate inflammation-induced DNA methylation changes using a mouse model in which inflammation is associated with cancer susceptibility. It has been hypothesized that the Polycomb repression complex present in stem cells at specific gene loci accelerates or facilitates recruitment of DNA methyltransferases and de novo methylation of CpG-rich sequences during the process of cell transformation. We will test this hypothesis using a neural stem cell and brain tumor model, which includes neural stem cells, stem cell-like glioma cells, and DNA samples from glioma tumors. Finally, the connection between Ras transformation, the Polycomb complex and DNA hypermethylation will be analyzed. NIH3T3 cells and immortalized epithelial cells will be transformed with the activated K-ras oncogene. We will determine if K-ras-induced transformation operates through the Polycomb complex to promote de novo methylation of CpG islands.